JP5986652B2 - Storage system using real storage area dynamic allocation method - Google Patents

Description

  The present invention relates to a method for dynamically allocating a real storage area to a control information area of a virtual volume in a storage apparatus.

  As disclosed in Patent Document 1, a storage apparatus is disclosed in which a plurality of microprocessors and a local memory accessible only from each microprocessor are provided. This storage apparatus grants ownership to a microprocessor that executes input / output processing with a storage device, and stores cache control information necessary for input / output processing with the storage device in the local memory of the microprocessor. .

  According to Patent Document 2, in the process related to the dynamic allocation of the real storage area to the virtual storage area, the allocation of the real storage area to the virtual storage area triggered by the host write to the virtual storage area, and the virtual storage Address conversion from the area to the real storage area is performed. The address conversion information required for input / output to the virtual storage area has a problem of increasing in proportion to the number and capacity of virtual volumes, but is stored in an inexpensive storage device, although access performance is inferior to that of local memory. Can be solved.

International Publication No. 2010/131373 US Patent No. 7613896

  If the microprocessor with ownership of the virtual storage device comprising the virtual storage area does not have ownership of the storage device that stores the address conversion information described above, address conversion required for input / output processing with the virtual storage device Unable to read information. Further, even if the owner right can be held, the volume is accessed across the owner right, so that there is a problem that the performance is lowered. Note that volume access across owner rights means, for example, changing a microprocessor having access rights to the volume (owner rights transfer). Furthermore, in the method of storing the address conversion information necessary for input / output to the virtual storage area in a distributed manner in the storage device having inferior access performance compared to the local memory, the address conversion on the storage device is performed when frequent access occurs. Access to information will also increase. Therefore, there is a problem that the overall access performance is deteriorated.

  It is an object of the present invention to store address conversion information from a virtual storage area to a real storage area on a storage device, to enable access from a microprocessor having ownership of the virtual storage device, and to improve access performance. .

In order to achieve the above object, the present invention provides:
(1) A first real storage area (hereinafter referred to as a page) is assigned from a plurality of storage devices to a first virtual storage area, and a page assigned to the first virtual storage area is assigned to a second virtual storage area. A second real storage area (hereinafter referred to as a track) is allocated. A page consists of a plurality of tracks, and stores track data in the order assigned to the second virtual storage area.
(2) The size of the first address conversion information between the first virtual storage area and the plurality of storage devices is proportional to the number of pages and is stored in the shared memory. The size of the second address conversion information between the second virtual storage area and the first virtual storage area is proportional to the number of tracks, and is stored on the page assigned to the first virtual storage area.
(3) By storing the second address conversion information necessary for input / output with the virtual storage device composed of the second virtual storage area on the page assigned to the first virtual storage area, The cache control information necessary for accessing the address translation information is stored in a local memory of a microprocessor having ownership of the virtual storage device.

  According to the present invention, the microprocessor can exclusively execute input / output processing for a virtual storage device that stores address translation information in the storage device.

1 is a block configuration diagram showing the overall configuration of a storage system. It is a block diagram which shows the structure of the microprogram performed within a storage controller. It is a block diagram which shows the logical structure in a storage apparatus. It is a block diagram which shows the management table preserve | saved in SM (Shared Memory). It is a block diagram of a management table for managing the correspondence between MPPK (Micro Processor PacKage) numbers and HDEV numbers. It is a block diagram of the management table for managing the correspondence of a HDEV (Host logical DEVice) number and a LDEV (Logical DEVice) number. It is a block diagram of a management table for managing the correspondence between HDEV numbers and volume sizes. It is a management table block diagram which manages the relationship between the HDEV number and the attribute of HDEV. It is a management table block diagram which manages the relationship between a HDEV number and a disk number. It is a management table block diagram which manages the relationship between a LDEV number and a pool number. It is a management table block diagram which manages the relationship between a pool number and a disk number. It is a block diagram of PMT (Page Mapping Table) used by a page allocation process. It is a block diagram of a PMT management directory. It is a block diagram of a TMT management directory. It is a block diagram of TMT (Track Mapping Table) used in the track allocation process in a page. It is a flowchart of the virtual volume creation processing in the first embodiment. It is a flowchart of the track | truck allocation process to the virtual volume in 1st Example. It is a management table block diagram which manages the relationship between the track number and copy status in 2nd Example. It is a management table block diagram which manages the relationship between the track number and update state in 2nd Example. It is a management table block diagram of the pair status in 2nd Example. It is a flowchart of the copy pair creation processing in the second embodiment. It is a flowchart of the background copy process by the copy job in 2nd Example. It is a flowchart for demonstrating the preceding copy process of the host write command opportunity in 2nd Example. It is a flowchart of differential restore command processing in the second embodiment. It is a flowchart of the track allocation release process triggered by the MF host write command in the third embodiment. It is a flowchart of the ownership transfer process in 4th Example. It is a flowchart of the volume role change in 3rd Example. It is a block block diagram which shows the whole structure of the storage system in 5th Example. It is a file structure figure in 5th Example.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, various types of information may be described using an expression such as “management table”, but the various types of information may be expressed using a data structure other than a table. Further, the “management table” can be referred to as “management information” to indicate that it does not depend on the data structure.

  Further, the process may be described with “program” as the subject. The program is executed by a processor, for example, an MP (Micro Processor), and performs a predetermined process. Note that the subject of the processing may be a processor because it is performed using storage resources (for example, a memory) and a communication interface device (for example, a communication port) as appropriate. The processor may have dedicated hardware in addition to the MP. The computer program may be installed on each computer from a program source. The program source may be, for example, a program distribution server or a storage medium.

  Each element, for example, LDEV (Logical DEVice) can be identified by a number or the like, but other types of identification information such as a name may be used as long as it is identifiable information. In the drawings and description of the present invention, the same reference numerals are given to the same parts, but the present invention is not limited to the present embodiment, and any application examples that meet the idea of the present invention are technical. Included in the range. Further, unless specifically limited, each component may be plural or singular.

  A dynamic real storage area allocation method to a control information area of a virtual volume in the storage apparatus according to the first embodiment of the present invention will be described with reference to FIGS. First, an outline of the hardware configuration and software configuration (program structure, table structure) will be described with reference to FIGS. 1 to 4, and then each management table will be described in detail with reference to FIGS. Finally, specific operations (processes) will be described with reference to FIGS.

  First, the overall configuration of the storage system will be described with reference to FIG. The storage system includes a plurality of storage apparatuses 101, a plurality of main frame host computers (hereinafter referred to as MF host computers) 102, and a management terminal computer 103.

  Each storage apparatus 101 and each MF host computer 102 are connected via a network 112. One of the plurality of storage apparatuses 101 and the management terminal computer 103 are connected via a network 111. Each MF host computer 102 is connected to the network 112 via a communication line 114. Each storage apparatus 101 is connected to the network 112 via the communication line 116. The management terminal computer 103 is connected to the network 111 via the communication line 113. The first storage device 101 is connected to the network 111 via the communication line 115.

  The communication lines 113 to 116 are configured as wires such as metal cables and optical fiber cables, for example. However, each MF host computer 102 and each storage device 101, each storage device 101 and the management terminal computer 103, and each MF host computer 102 and the management terminal computer 103 can be connected wirelessly. In this case, the communication lines 113 to 116 are not necessary. Further, the network 111 and the network 112 may be a common network. Each of the networks 111 and 112 is a communication network, for example, a SAN (Storage Area Network) or a LAN (Local Area Network).

  Next, the configuration of each storage apparatus 101 will be described. Each storage device 101 includes one or more storage controllers 131 and one or more disks 132. One or more disks 132 are at least SSD (Solid State Drive), SAS (Serial Attached SCSI) type HDD (Hard Disk Drive), and SATA (Serial Advanced Technology Attachment) type HDD as storage devices for storing data. It is included. The disk 132 may be another type of physical storage device such as a write-once optical disk medium instead of or in addition to at least one of them.

  One or more disks 132 are connected to the storage controller 131 via a communication line such as a fiber channel cable. A plurality of disks 132 can constitute one or a plurality of RAID (Redundant Array of Independent Disks) groups.

  Next, the configuration of the storage controller 131 will be described. The storage controller 131 controls the data input / output processing with respect to the disk 132, that is, the writing (reading) and reading (reading) of data to the disk 132, with the command received from each MF host computer 102.

  The storage controller 131 is a logical device in which a real storage area has already been allocated to the access request source MF host computer 102, or a logical device configured by a virtual storage area used in the thin provisioning function described later. Are provided as a logical device or logical volume to be accessed.

  For a logical device composed of virtual storage areas, the real storage area may be already allocated or unallocated. At this time, the storage controller 131 can refer to or identify a real storage area or a virtual storage area by a cylinder / head number (hereinafter referred to as a track number). The storage controller 131 includes a storage resource, a communication interface device (hereinafter, the interface device is abbreviated as “I / F”), and an arithmetic device connected thereto.

  The arithmetic unit is an MP (Micro Processor) 141 in an MPPK (Micro Processor PacKage) 123 or the like. The storage resources include an LM (Local Memory) 142 in the MPPK 123, a shared memory SM (Shared Memory) 143 in a CMPK (Cache Memory PacKage) 124, a cache memory CM (Cache Memory) 144, and the like.

  The communication I / F includes a CHAPK (Channel Adapter PacKage) 121 that receives read and write from the MF host computer 102, a NIC (Network Interface Card) 125 that receives a local copy instruction from the management terminal computer 103, and data from the disk. And a DKAPK (DisK Adapter PacKage) 122 for receiving and transmitting the message.

  These CMPK 124, MPPK 123, CHAPK 121, NIC 125, and DKAPK 122 are connected to each other via a communication line such as a bus and a SWPK (Switch PacKage) 126 in charge of switching communication.

  The hardware configuration of each MF host computer 102 and management terminal computer 103 is a general computer having a communication interface device, a storage resource, and an arithmetic device connected thereto. Examples of the communication interface device include a host bus adapter (HBA) for performing communication via the network 112 and a NIC for performing communication via the network 111. The storage resource includes a memory, a built-in HDD, and the like.

  FIG. 2 shows the configuration of the microprogram 201 executed in the storage controller 131. One or a plurality of microprograms 201 are read into the local memory LM142 of FIG. The microprogram 201 includes a command control unit 211, a RAID control unit 212, and an ownership control unit 213. When the MP 141 executes the control units 211 to 213 in the microprogram 201 read into the LM 142, various processes described later are performed.

  The CM control information 202 controls information cached on the LM 142. The management table cache 203 stores (caches) each management table (FIG. 4) on the shared memory SM143 in the local memory LM142. Thereby, the MP 141 can acquire the information of each management table from the LM 142 that can be accessed at high speed instead of the SM 143 that takes time to access. The CM 144 temporarily stores the write data received from each MF host computer 102 and the data read from the disk 132 by the RAID control unit 212 in the microprogram 201.

  The ownership control unit 214 is in charge of setting and referring to the management table 401 that manages the correspondence between the MPPK number having the ownership of the HDEV described later and the HDEV number. Then, in response to a setting or reference inquiry from the command control unit 211 or CHAPK 121, the correspondence relationship between the MPPK number having the ownership of the HDEV and the HDEV number is set or referred to. Further, the ownership control unit 214 sets a correspondence relationship between the MPPK number having the ownership of the HDEV and the HDEV number when a change of ownership is requested from the command control unit in the ownership transfer processing described later. To do. Thereafter, the old information is invalidated and updated to the latest information for the copy of the management table 401 in the CHAPK 121 and on the LM 142.

  The CHAPK 121 is connected to each MF host computer 102 via the network 112, receives an access command (write command or read command) as an access request, and transfers the received access command to the command control unit 211. The CHAPK 121 identifies the MPPK having the ownership of the HDEV based on the HDEV number that is the parameter of the read or write command in the transfer process and the copy of the management table 401 in the CHAPK 121. Thereafter, the CHAPK 121 transfers an access command to the command control unit 211 in the identified MPPK in order to process a read or write command.

  When the NIC 125 is connected to the management terminal computer 103 via the network 111 and receives a data copy instruction between volumes described later, the NIC 125 transfers the received instruction to the command control unit 211. The DKAPK 122 performs data transmission / reception between each disk 132 and storage resources (LM 142, SM 143, and CM 144). The DKAPK 122 is connected to each disk 132 via a communication path.

  Next, the basic operation of the storage apparatus 101 will be described. When the storage controller 131 receives a write command from any MF host computer 102 via the CHAPK 121, the storage controller 131 stores the write data received from the MF host computer 102 in the CM 144.

  The storage controller 131 writes the write data stored in the CM 144 to the disk 132 via the DKAPK 122. The storage controller 131 notifies the MF host computer 102 of the completion of the write command processing when it is stored in the CM 144 or when write data is written to the disk 132.

  When a read command is received from the MF host computer 102, the storage controller 131 checks whether data (read target data) specified by the parameter in the read command is stored in the CM 144.

  When the read target data is stored in the CM 144, the storage controller 131 reads the read target data from the CM 144 and transmits the read read target data to the MF host computer 102 via the CHAPK 121. On the other hand, when the read target data is not stored in the CM 144, the storage controller 131 reads the read target data from one or a plurality of disks 132 via the DKAPK 122, and stores the read read target data in the CM 144. Thereafter, the storage controller 131 transmits the read target data stored in the CM 144 to the MF host computer 102 via the CHAPK 121.

  Next, the logical configuration of the storage apparatus 101 will be described with reference to FIG. The storage apparatus 101 includes one or a plurality of host logical devices (hereinafter referred to as “HDEV”) 311 as host logical devices constituting a storage area referred to by each MF host computer 102 or management terminal computer 103. Have.

  The HDEV 311 is assigned a unique HDEV number in the storage apparatus 101, and the MF host computer 102 and the management terminal computer 103 identify the HDEV 311 by the HDEV number. For example, an OS (Operating System) 301 on each MF host computer 102 performs read access or write access to the HDEV 311.

  The HDEV 311 is composed of a set of tracks (track 1, track 2, track 3...). The cylinder head number (track number) assigned to each track is a reference object when each MF host computer 102 or management terminal computer 103 identifies each track. The HDEV 311 is classified into a normal volume and a virtual volume according to the management table 404 that manages the HDEV number and volume attribute shown in FIGS.

  When the HDEV 311 is a normal volume, it is defined as a storage area in one or a plurality of disks 132. Further, the HDEV 311 may be defined as a storage area composed of a plurality of RAID groups. When the HDEV 311 is a virtual volume, the track data in the HDEV 311 is stored on one or more logical devices (hereinafter sometimes referred to as LDEV).

  The LDEV 312 is given a unique LDEV number in the storage apparatus 101, and the microprogram 201 identifies the LDEV 312 by the LDEV number. The LDEV 312 is composed of a plurality of tracks corresponding to track numbers. Each track is composed of one or a plurality of records (not shown), and data read or written by the MF host computer 102 is stored in each record.

  The track in the LDEV 312 stores track data corresponding to the track number in the HDEV 311 described above. The correspondence between the track numbers in the HDEV and the track numbers in the LDEV is managed by a TMT (Track Mapping Table) 410 in FIGS.

  At this time, all the track sizes in the storage apparatus 101 are constant values. Each disk 132 is given a unique disk number in the storage apparatus 101, and the microprogram 201 identifies each disk 132 by this disk number.

  The LDEV 312 is stored on a virtual volume used in the thin provisioning function. A virtual volume in which LDEV 312 data is stored is associated with a pool 313 that provides a real storage area for the virtual volume. The pool 313 is given a unique pool number in the storage apparatus 101, and the microprogram 201 identifies the pool 313 by this pool number.

  The pool 313 is composed of one or a plurality of disks 132. The pool 313 may be configured by one or a plurality of RAID groups. One or a plurality of disks 132 may be inside or outside the storage apparatus.

  FIG. 4 shows the configuration of each management table stored in the SM 143. Each management table is created when the storage apparatus 101 is started or dynamically created as necessary. In addition, each field in the management tables 401 to 410 is updated when the logical configuration of the storage apparatus 101 is changed.

  FIG. 5 shows a management table 401 for managing the correspondence between the MPPK number having the ownership of the HDEV and the HDEV number. Ownership defines MPPK for processing a read or write request from the host to the HDEV.

  The MPPK 123 having ownership can exclusively handle control information necessary for processing a read or write request from the host for the HDEV 311 from other MPPKs 123. This control information is information indicating whether read or write data for the HDEV 311 is cached in the CM 144, PMT (Page Mapping Table) when the HDEV shown in FIGS. 12 and 13 has a virtual volume attribute, and the like. A table 401 for managing the correspondence between MPPK numbers and HDEV numbers is stored in the SM 143 or the like.

  FIG. 6 shows the configuration of a table 402 that manages the correspondence between the HDEV number and the LDEV number when the HDEV 311 is configured from a virtual volume. The management table 402 includes an HDEV number field 601 and an LDEV number field 602.

  The HDEV number is a number for uniquely identifying the HDEV 311 within the storage apparatus 101, and each entry in the HDEV number field 601 stores a number corresponding to the HDEV 311. The LDEV number is a number for uniquely identifying the LDEV 312 in the storage apparatus 101, and a number corresponding to the LDEV 312 is stored in each entry of the LDEV number field 602. A management table 402 for managing the correspondence between HDEV numbers and LDEV numbers is stored in the SM 143 or the like.

  FIG. 7 shows the configuration of the management table 403 for managing the correspondence between the HDEV number and the volume size. The management table 403 includes an HDEV number field 701 and a volume size field 702. The HDEV number is a number for uniquely identifying the HDEV 311 in the storage apparatus 101, and each entry in the HDEV number field 701 stores a number corresponding to the HDEV 311.

  Each entry of the volume size field 702 stores a volume size corresponding to the HDEV 311. The volume size is the maximum capacity that can store data in the normal volume or virtual volume, and each entry in the volume size field 702 stores the capacity of the normal volume or virtual volume as a numerical value. A management table 403 for managing the correspondence between HDEV numbers and volume sizes is stored in the SM 143 or the like.

  FIG. 8 shows the configuration of the management table 404 for managing the correspondence between HDEV numbers and volume attributes. The management table 404 includes an HDEV number field 801 and a volume attribute field 802. The HDEV number is a number for uniquely identifying the HDEV 311 in the storage apparatus 101, and each entry in the HDEV number field 801 stores a number corresponding to the HDEV 311.

  The volume attribute specifies that the volume attribute of the HDEV 311 is a normal volume or a virtual volume, and each entry in the volume attribute field 802 stores the name of the normal volume or the virtual volume.

  When the volume attribute of the HDEV 311 is a normal volume, this HDEV 311 indicates that the volume is composed of real storage areas in one or more disks 132. A management table 404 for managing the correspondence between HDEV numbers and volume attributes is stored in the SM 143 or the like.

  FIG. 9 shows the configuration of the management table 405 for managing the correspondence between the HDEV number and the disk number when the HDEV 311 is a normal volume. The management table 405 includes an HDEV number field 901 and a disk number field 902. The HDEV number is a number for uniquely identifying the HDEV 311 in the storage apparatus 101, and each entry in the HDEV number field 901 stores a number corresponding to the HDEV 311.

  The disk number is a number for uniquely identifying the disk 132 in the storage apparatus 101, and the number of the disk 132 constituting the HDEV 311 is stored in each entry of the disk number field 902. A management table 405 for managing the correspondence between HDEV numbers and disk numbers is stored in the SM 143 or the like.

  FIG. 10 shows the configuration of a management table 406 that manages the correspondence between LDEV numbers, pool numbers, and PMT (Page Mapping Table) addresses in SM143. The management table 406 is a management table when the volume attribute of the LDEV 312 is a virtual volume. The management table 406 includes an LDEV number field 1001, a pool number field 1002, and a PMT address field 1003 in the SM 143.

  The LDEV number is a number for uniquely identifying the LDEV 312 in the storage apparatus 101, and each entry in the LDEV number field 1001 stores a number corresponding to the LDEV 312. The pool number is a number for uniquely identifying the pool 313 within the storage apparatus 101, and each entry in the pool number field 1002 stores the number of the pool 313 corresponding to the LDEV 312. Each entry in the PMT address field 1003 in the SM stores a PMT address in the SM 143 (described later) corresponding to the LDEV 312. A management table 406 for managing the correspondence between LDEV numbers and pool numbers is stored in the SM 143 or the like.

  FIG. 11 shows the configuration of the management table 407 for managing the correspondence between pool numbers and disk numbers. The management table 407 is a management table when the volume attribute of the HDEV 311 is a virtual volume, and includes a pool number field 1101 and a disk number field 1102.

  The pool number is a number for uniquely identifying the pool 313 within the storage apparatus 101, and each entry in the pool number field 1101 stores a number corresponding to the pool 313. The disk number is a number for uniquely identifying the disk 132 in the storage apparatus 101, and each entry of the disk number field 1102 stores the number of the disk 132 to which the real storage area is allocated to the pool 313. . A management table 407 for managing the correspondence between pool numbers and disk numbers is stored in the SM 143 or the like.

  In this embodiment, the HDEV 311 is managed as a virtual volume having a virtual storage area, and when accessing the virtual volume, a first real storage area (hereinafter referred to as a page) is allocated from the pool to the virtual volume. A second real storage area (track) is allocated from the real storage area (page).

  Real storage area allocation processing to a virtual volume includes page allocation processing and track allocation processing from a page. In the page allocation process, the PMT (Page Mapping Table) 408 and the PMT management directory 409 allocate a real storage area to the LDEV in units of pages in response to a page allocation request to the LDEV. This function of allocating pages to the LDEV is called a thin provisioning function.

  FIG. 12 shows the configuration of a PMT (Page Mapping Table) 408 used in the page assignment processing to the LDEV. The PMT 408 includes a page number field 1201, a page type field 1202, a page head address field 1203 in the LDEV, a disk number field 1204, a page head address field 1205 in the disk, and a page assignment determination information field 1206. The PMT 408 is provided for each LDEV 312 and stored in the LM 142 or the like. Alternatively, the information can be stored in the SM 143 and cached in the LM 142.

  The page number 1201 in the LDEV of the PMT 408, the page type 1202 corresponding to the page number 1201 in the LDEV, the page head address 1203 in the LDEV, the disk number 1204, the page head address 1205 in the disk, and the page allocation determination information 1206 The combination is called a page mapping entry.

  Each entry of the page number field 1201 stores the page number of the page assigned in the LDEV 312. A page type field 1202 represents the type of information stored in a page assigned in the LDEV 312. The page type is a user data page in which track data read or written from the MF host 102 is stored, or a control information page in which a TMT (Track Mapping Table) 410 is stored.

  Each entry in the page head address field 1203 in the LDEV stores the head address of the page assigned in the LDEV 312. Each entry of the disk number field 1204 stores the number of the disk 132 constituting the LDEV 312.

  The page head address 1205 in the disk is the head address of the real storage area in the disk, and the head address of the disk 132 constituting the LDEV 312 is stored in each entry of the page head address field 1205 in the disk. Each entry of the page allocation determination information field 1206 stores “allocated” information when a page is allocated to the LDEV 312 and “unallocated” information when it is not allocated. That is, it is possible to manage the allocation state of the real storage area to the LDEV on a page basis by this page allocation determination information.

  A page is a storage area of a certain size stored in the pool 313, and is composed of one or a plurality of tracks as a unit for dividing and managing the storage area of the LDEV 312. A track is a fixed size managed by the storage apparatus 101, and is composed of, for example, 59,392 bytes.

  FIG. 13 shows the configuration of a PMT (Page Mapping Table) management directory 409 that manages the correspondence between the LDEV 312 and the PMT 408. The PMT management directory 409 includes an LDEV number field 1301 and a PMT address field 1302 and is stored in the SM 143 or the like. Each entry of the LDEV number field 1301 stores the LDEV 312 number. Each entry of the PMT address field 1302 stores the address of the PMT 408 for managing the page of the LDEV 312.

  FIG. 14 shows a configuration of a TMT (Track Mapping Table) management directory 410 that manages allocation information from the LDEV 312 to the HDEV 311. The TMT management directory 410 includes an LDEV number field 1401, an initial data track number field 1402 in the LDEV, and a next assigned track number field 1403 in the LDEV, and is stored in the SM 143 or the like.

  Each entry in the LDEV number field 1401 stores the LDEV 312 number. Each entry of the initial data track number field 1402 stores the track number of the initial data track on the LDEV 312. Each entry of the next allocation track number field 1403 stores a track number that is a candidate for the next new allocation on the LDEV 312.

  FIG. 15 shows a configuration of a TMT (Track Mapping Table) 1511 used when executing a process of assigning a track to the HDEV 311 from a page assigned to the LDEV 312. The TMT 1511 includes a track number field 1501 in the HDEV, a track number field 1502 in the LDEV, and a track assigned determination information field 1503. Note that the TMT 1511 is set for each HDEV 311 and stored on one or more pages in the LDEV 312.

  A combination of the track number 1501 in the HDEV, the track number 1502 in the LDEV corresponding to the track number 1501 in the HDEV, and the track assignment determination information 1503 in the TMT 1511 is referred to as a track mapping entry. In each entry of the track number field 1501 in the HDEV, the track number of the track assigned to the HDEV 311 from the page assigned to the LDEV 312 is stored. In each entry of the track number field 1502 in the LDEV, the track number in the LDEV 312 assigned to the HDEV 311 is stored.

  Each entry of the track assignment determination information field 1503 stores “assigned” information when the track has been assigned to the HDEV 311 and “unassigned” information when the track is not assigned. That is, with this allocation determination information, the allocation state of the real storage area from the LDEV to the virtual storage area can be managed in track units.

  Next, the operation of this embodiment will be described with reference to FIGS. First, a process in which the command control unit 211 creates a virtual volume attribute HDEV upon reception of a virtual volume creation command from the management terminal computer 103 will be described. The parameters of the virtual volume creation command are, for example, an HDEV number (16-a) and a virtual volume size (16-b). The page size is, for example, 59,392 × 672 bytes.

  When receiving the parameters of the virtual volume creation command from the management terminal computer 103 (S1601), the command control unit 211 uses the PMT (Page Mapping) with the virtual volume size (16-b) and the page size (59,392 × 672 bytes). Table) The SM size (16-c) required for creation is calculated.

  If the size of the unused area on the SM 143 in the CMPK 124 is less than the SM size (16-c) (S1608: “Yes”), the command control unit 211 notifies the management terminal computer 103 that the volume cannot be created. A response is made (S1602). If the size of the unused area is equal to or larger than the SM size (16-c) (S1608: “No”), the command control unit 211 secures the PMT 408 in the unused area on the SM 143, and this PMT address (16-d). To get.

  The command control unit 211 calculates the size (16-e) of a TMT (Track Mapping Table) necessary for the virtual volume based on the virtual volume size (16-b) and the track size. The command control unit 211 calculates the number (16-f) of one or a plurality of pages (control information pages) storing the TMT based on the TMT size (16-e) and the page size (S1603). .

  If the pool capacity is insufficient for the number of control information pages required for the virtual volume (S1609: “Yes”), the command control unit 211 responds to the management terminal computer 103 that volume creation is not possible (S1604). . If the control information page necessary for the virtual volume can be secured in the pool (S1609: “No”), the command control unit 211 determines that one or more from the pool to the virtual volume based on the number of control information pages (16-f). Allocate multiple control information pages. The command control unit 211 initializes one or more assigned control information pages. The initial value is, for example, “0” (S1605).

  The command control unit 211 sets the MPPK number having the ownership of this volume and the HDEV number (16-a) in the unused entry of the management table 401 (FIG. 5) that manages the correspondence between the MPPK number and the HDEV number. To do. The MPPK number (16-a) having ownership for the volume may be specified by a parameter of the virtual volume creation command. Further, the command control unit 211 may set the MPPK number having the smallest load in the storage apparatus as the MPPK number (16-a) having ownership over the volume.

  The command control unit 211 adds the HDEV number (16-a) and the unused LDEV number (16-g) to the unused entry in the management table 402 (FIG. 6) that manages the correspondence between the HDEV number and the LDEV number. Set.

  The command control unit 211 sets the HDEV number (16-a) and the volume size (16-b) in the unused entry of the management table 403 (FIG. 7) that manages the correspondence between the HDEV number and the volume size.

  The command control unit 211 sets the HDEV number (16-a) and the “virtual volume” attribute in the unused entry of the management table 404 (FIG. 8) that manages the correspondence between the HDEV number and the attribute of the HDEV.

  The command control unit 211 stores the LDEV number (16-g) and the LDEV number (in the unused entry in the management table 406 (FIG. 10) that manages the correspondence between the LDEV number (16-g), the pool number, and the PMT address. A pool number and a PMT address (16-d) for page allocation are set in the 16-g) LDEV.

  Further, the command control unit 211 performs a new track assignment process from the LDEV with the LDEV number (16-g), and obtains the track number (16-h) of the assigned track and the LDEV internal assigned track number (16-i). get. Thereafter, the LDEV number (16-g), the initial data track number in LDEV (16-h), and the next assigned track number in LDEV (16-i) are added to the unused entries of TMT (Track Mapping Table) 410 (FIG. 14). Are set (S1606). Finally, the command control unit 211 returns a volume creation completion response to the management terminal computer 103 (S1607).

  Next, processing when the command control unit 211 receives a write access from the MF host computer 102 will be described with reference to FIG. The parameters at the time of write access are, for example, the HDEV number (17-a) and the track number (17-b) in the HDEV. The page size is, for example, 59,392 × 672 bytes (S1701).

  When receiving a write access parameter from the MF host computer 102, the command control unit 211 calculates an address (17-c) in the HDEV 311 based on the track number (17-b) and the track size of the HDEV 311.

  The command control unit 211 acquires the LDEV number (17-d) corresponding to the HDEV 311 from the management table 402 (FIG. 6) that manages the correspondence between the HDEV number (17-a) and the HDEV number and the LDEV number. . The command control unit 211 acquires the PMT 408 (FIG. 13) corresponding to the LDEV number (17-d) based on the LDEV number (17-d) and the PMT management directory 409. The command control unit 211 examines one or a plurality of page mapping entries whose page type 1202 is the control information page in the PMT 408, and acquires the position of the TMT 1511 from the page head address 1203 in the LDEV (S1702).

  The command control unit 211 acquires the track mapping entry (17-e) corresponding to the track number (17-a) based on the position of the TMT 1511 and the track number (17-a).

The command control unit 211 refers to the track allocation determination information 1503 in the track mapping entry (17-e), and when the track allocation determination information 1503 is “allocated” (S1705: “No”), The operations (W1) to (W6) are performed.
(W1) The command control unit 211 acquires the track number (17-f) in the LDEV in the track mapping entry (17-e).
(W2) The command control unit 211 calculates the page number (17-g) in the LDEV including the track number (17-f) from the track number (17-f) and the page size, and the start address in the page. Calculate the offset (17-h).
(W3) The command control unit 211 sets the page mapping entry (17-i) corresponding to the page number (17-g) in the LDEV based on the PMT 408 and the page number (17-g) in the LDEV. get.
(W4) The command control unit 211 obtains the page head address (17-j) on the disk from the page mapping entry (17-i), adds the offset (17-h), and writes it on the disk of the write target track. Address (17-k) is calculated.
(W5) The command control unit 211 writes the data sent from the MF host computer 102 to the address (17-k) (S1703).
(W6) A write command completion response is returned to the MF host computer (S1704).

  The command control unit 211 refers to the track allocation determination information in the track mapping entry (17-e). When the track allocation determination information is “unallocated” (S1705: “Yes”), the command control unit 211 performs the following operation. Do.

  First, the command control unit 211 sets the track number (17-l) in the LDEV indicated by the next assigned track address 1403 of the TMT management directory 410 (FIG. 14) to the track number in the track entry (17-e) in the TMT. Set.

  The command control unit 211 further sets the track allocation determination information to “allocated”. Then, the value of the next allocation track address 1403 in the TMT management directory 410 is incremented to be the next track allocation candidate destination (S1706 to S1709). Thereafter, the command control unit 211 performs the processes (W1) to (W6). The processes (W1) to (W6) are referred to as “S1711”, and are used in the following description.

  When the page allocation determination information of the page mapping entry (17-i) acquired in (W3) is “unallocated” (S1706: “Yes”), the command control unit 211 determines the LDEV number, the pool number, and the PMT address. The pool number (17-m) corresponding to the LDEV number (17-d) is acquired with reference to the management table 406 (FIG. 10) that manages the correspondence with

  The command control unit 211 corresponds to the pool number (17-m) by referring to the management table 407 (FIG. 11) that manages the correspondence between the pool number and the disk number based on the acquired pool number (17-m). The obtained disk number (17-n) is acquired.

  Based on the acquired disk number (17-n), the command control unit 211 allocates a real storage area in units of pages from one or more disks 132 belonging to the pool 313 corresponding to the LDEV number (17-d). Is performed (S1707).

  At this time, the command control unit 211 searches the page entry including the track indicated by the write access destination address with reference to the PMT 408 (FIG. 12), registers the head address of the newly assigned page in the PMT 408, and has already been assigned. The “unassigned” in the determination information field 1206 is set to “assigned”, and the PMT 408 is updated. The above is the processing when the command control unit 211 receives a write access from the MF host computer 102.

  At the time of read access from the MF host computer 102, the address (17-k) of the read target track is calculated from the above (W1) to (W4), and the data at the address (17-k) is transmitted to the MF host computer 102. To do.

  The command control unit 211 refers to the track allocation determination information in the track mapping entry (17-e). When the track allocation determination information is “unallocated”, the initial data on the TMT management directory 410 (FIG. 4) The track having the track number 1402 (FIG. 14) can be read from the MF host computer 102.

  The above is the process of assigning pages from the pool to the LDEV and assigning tracks from the LDEV to the HDEV for the virtual volume. As a result, it is possible to access the track mapping information in the hierarchical memory (FIG. 3) of the distributed memory type system (FIG. 1) so that the owner right does not straddle. For this reason, access to high-speed track mapping information, that is, access to a high-speed volume can be realized.

  The data copy processing from the normal volume to the virtual volume according to the second embodiment of the present invention will be described with reference to FIGS. Specifically, the copy source HDEV 311 (primary volume) is managed as a normal volume composed of real storage areas, and the copy destination HDEV 311 (secondary volume) is managed as a virtual volume composed of virtual storage areas, and the primary volume ( Data copy is performed between 18-a) and the secondary volume (18-b).

  FIG. 18 shows the configuration of a management table (copy BM table) 1811 for managing track numbers and their copy states. The copy BM table 1811 includes a track number field 1801 and a copy status field 1802. The track number is a number for uniquely identifying a track in the HDEV 311, and each entry in the track number field 1801 stores a number corresponding to each track in the HDEV 311.

  The copy status specifies that the track is “copied” or “not copied”, and each entry in the copy status field 1802 includes “copied” (for example, the bit is set to “b′1”). ") Or" uncopied "(for example, the bit is off" b'0 ") is stored.

  FIG. 19 shows the configuration of a table (difference BM table) 1911 for managing track numbers and their update states. The difference BM table 1911 includes a track number field 1901 and an update state field 1902.

  The track number is a number for uniquely identifying a track in the HDEV 311, and each entry in the track number field 1901 stores a number corresponding to each track in the HDEV 311.

  The update status specifies that the track is “updated” or “not updated”, and each entry in the update status field 1902 includes “updated” (for example, the bit is set to “b′1”). ") Or" not updated "(for example, the bit is on" b'0 ") is stored.

  FIG. 20 shows the configuration of the pair management table 2011 used when executing copy processing between HDEVs. The pair management table 2011 includes a pair number field 2001, a copy source HDEV number field 2002, a copy source track range field 2003, a copy destination HDEV number field 2004, a copy destination track range field 2005, a copy BM table address field 2006, and a differential BM table address. Field 2007, set for each copy pair, and stored in SM 143 or the like.

  Each entry in the pair number field 2001 of the pair management table 2011 stores a pair number assigned for each pair. Each entry in the copy source HDEV number field 2002 stores the HDEV number of the copy source HDEV. Each entry in the copy source track range field 2003 stores the copy start track number and copy end track number of the copy source HDEV.

  Each entry in the copy destination HDEV number field 2004 stores the HDEV number of the copy destination HDEV. Each entry in the copy destination track range field 2005 stores a copy start track number and a copy end track number of the copy destination HDEV. Each entry of the copy BM table address field 2006 stores the address of the copy BM table. Each entry of the differential BM table address field 2007 stores an address of the differential BM table.

  Next, a pair creation process in data copy from the normal volume attribute HDEV 311 (primary volume) to the virtual volume attribute HDEV 311 (secondary volume) will be described with reference to FIG. When receiving a copy pair creation command between the primary volume (21-a) and the secondary volume (21-b) from the management terminal computer 103, the command control unit 211 performs the following operations.

  The command parameters include, for example, the HDEV number (21-c) of the primary volume (21-a), the HDEV number (21-d) of the secondary volume (21-b), and the copy source track of the primary volume (21-a). The range (21-e) and the copy destination track range (21-f) of the secondary volume (21-b). The number of tracks in the copy source track range (21-e) is equal to the number of tracks in the copy destination track range (21-f) (S2101).

  The command control unit 211 calculates the SM143 capacity required for the copy BM table 1811 and the difference BM table 1911 from the copy source track range (21-e) or the copy destination track range (21-f) (S2102). If the required SM 143 capacity cannot be secured in the copy BM table 1811 and the difference BM table 1911 (S2107: “Yes”), the command control unit 211 responds to the management terminal computer 103 with a pair creation failure (S2103). .

  If the required SM143 capacity can be secured in the copy BM table 1811 and the difference BM table 1911 (S2107: “No”), the command control unit 211 secures the copy BM table 1811 and the difference BM table 1911 on the SM143. The command control unit 211 initializes the secured copy BM table 1811 and the difference BM table 1911. A typical initial value is “0”, but is not limited to this (S2104).

  The command control unit 211 newly assigns an unused pair number in the pair management table 2011. Then, the command control unit 211 copies the HDEV number (21-c) of the primary volume (21-a), the HDEV number (21-d) of the secondary volume (21-b), and the copy of the primary volume (21-a). A pair entry (21) includes an original track range (21-e), a copy destination track range (21-f) of the secondary volume (21-b), an address of the copy BM table 1811, and an address of the differential BM table 1911. -G) (S2105).

  The command control unit 211 returns a pair creation command completion response to the management terminal computer 103 (S2106). The above is the operation of the pair creation process in the data copy from the HDEV 311 with the normal volume attribute to the HDEV 311 with the virtual volume attribute.

  Next, a copy job operation in data copy from the normal volume attribute HDEV 311 (primary volume) to the virtual volume attribute HDEV 311 (secondary volume) will be described with reference to FIG.

  The command control unit 211 activates a copy job (not shown) after performing the pair creation process described above. The copy job performs the data copy process as follows based on the pair entry (21-g) set in the pair creation process described above (S2201).

The copy job refers to the pair entry (21-g) and sets the copy target track number (21-h) as the first track number in the copy source track range of the primary volume (21-a) (S2202). The copy job is processed in the following order (C1) to (C7) until the copy target track number (21-h) reaches the end track number of the copy source track range (S2203: “Yes”).
(C1) The copy job acquires the copy BM table 1811 from the copy BM table address in the pair entry (21-g).
(C2) The copy job includes an offset from the copy start track number (21-i) from the copy target track number (21-i) and the copy start track number (21-i) of the primary volume (21-a). 21-j) is calculated.
At this time, the difference between the track numbers (21-h) and (21-i) is equal to the offset (21-j).
(C3) The copy job acquires the offset (21-j) from the copy start track number (21-i) and the copy status of the track number (21-h) from the copy BM table 1811.
(C4) If the copy status of the track number (21-h) is “copied” (“No” in S2204), the next job (C5) to (C7) is not performed. The copy target track number (21-h) is incremented (S2208), and the process returns to (C1).
(C5) If the copy status of the track number (21-h) is “not copied” (“Yes” in S2204), the copy job has a secondary volume (21-b) of the pair entry (21-g). The copy destination track number (21-l) in the secondary volume (21-b) is calculated from the copy start track number (21-k) and the offset (21-j) in (2205).
(C6) The copy job copies the track data of the track number (21-h) of the primary volume (21-a) to the copy destination track number (21-l) of the secondary volume (21-b) (S2206). . The command control unit 211 refers to the TMT 1511, and if the allocation determination information of the copy destination track number (21-l) is “unallocated”, the copy destination track number (21-l) of the secondary volume (21-b). The above-described new track allocation to () (S1711: (W1) to (W6)) is performed.
(C7) The copy job sets the copy status of the copy BM table corresponding to the track number (21-h) of the primary volume (21-a) to “copied” (S2207).
The above is the operation of the copy job in the data copy from the HDEV 311 having the normal volume attribute to the HDEV 311 having the virtual volume attribute.

  Next, the preceding copy operation by writing from the MF host computer 102 in the data copy from the normal volume attribute HDEV 311 (primary volume) to the virtual volume attribute HDEV 311 (secondary volume) will be described with reference to FIG.

  When the command control unit 211 receives a write command for the primary volume (23-a) from the MF host computer 102 (S2301), the following operations (CW1) to (CW8) are performed. At this time, the command parameters are, for example, the HDEV number (23-b) of the primary volume (23-a) and the track number (23-c) of the write target track.

The command control unit 211 acquires one or more pair entries whose copy source HDEV number is (23-b) from the pair management table 2011. The command control unit 211 performs the following operation for each pair entry.
(CW1) The command control unit 211 checks whether the track number (23-c) of the write target track is included in the copy source track range in the pair entry. If not included, the processes (CW2) to (CW8) are skipped and nothing is done.
(CW2) The command control unit 211 acquires the copy BM table 1811 from the copy BM table pointer in the pair entry.
(CW3) The command control unit 211 calculates from the copy start track number from the track number (23-c) of the primary volume (23-a) and the copy start track number (23-d) in the copy source HDEV of the pair entry. Calculate the offset (23-e). At this time, the difference between the track numbers (23-c) and (23-d) is equal to the value of (23-e).
(CW4) The command control unit 211 acquires the offset (23-e) from the copy start track number and the copy status of the track number (23-c) from the copy BM table 1811.
(CW5) If the copy state of the track number (23-c) is “copied” (S2302: “No”), the command control unit 211 does not perform the processes (CW6) and (CW7), Process (CW8) is performed.
(CW6) If the copy state of the track number (21-c) is “not copied” (S2302: “Yes”), the command control unit 211 copies the copy start track number (23 in the copy destination HDEV of the pair entry) The copy destination track number (23-g) in the copy destination HDEV is calculated from -f) and the offset (23-e) (S2303).
(CW7) The command control unit 211 copies the pre-update track data of the track number (23-c) of the primary volume (23-a) to the copy destination track number (23-g) of the secondary volume (S2304). The command control unit 211 sets the copy status of the copy BM table 1811 corresponding to the track number (23-c) of the primary volume (23-a) to “copied” (S2305). In addition, the command control unit 211 sets the copy status of the differential BM table 1911 corresponding to the track number (23-c) of the primary volume (23-a) to “updated” (S2306). The command control unit 211 refers to the TMT 1511, and if the allocation determination information of the copy destination track number (23-g) is “unallocated”, the command to the copy destination track number (23-g) of the secondary volume is described above. New track allocation (S1711: (W1) to (W6)) is performed.
(CW8) The command control unit 211 writes the data received from the MF host computer 102 to the track number (23-c) of the primary volume (S2307).
Finally, the MF host computer 102 is notified of the completion of the write command. The write access from the MF host computer 102 to the primary volume (23-a) is performed in the data copy between the HDEV 311 (primary volume) with the normal volume attribute and the HDEV 311 (secondary volume) with the virtual volume attribute. This is the operation.

  Even when a write request from the MF host computer 102 to the secondary volume occurs during data copying between the normal volume attribute HDEV 311 (primary volume) and the virtual volume attribute HDEV 311 (secondary volume), the above (CW1) The write operation can be easily realized by executing the processing of CW8).

  Next, in the data copy between the normal volume attribute HDEV 311 (primary volume) (24-a) and the virtual volume attribute HDEV 311 (secondary volume) (24-b), the secondary volume (24-b) to the primary volume The differential restore operation to (24-a) will be described with reference to FIG.

  When the command control unit 211 receives a differential restore command for the pair entry with the pair number (24-c) from the MF host computer 102 (S2401), the following operation is performed. At this time, the command parameter is, for example, a pair number (24-c).

From the pair management table 2011, the command control unit 211 copies the copy source HDEV number (24-d) corresponding to the pair number (24-c), the copy source track range (24-e), and the copy destination HDEV number (24 -F), the copy destination track range (24-g), and the difference BM table are acquired. Further, the command control unit 211 sets the restoration target track (24-h) as the copy destination track start number (S2402) and continues to the next (CR1) until the copy destination track end number is reached (S2403: “Yes”). ) To (CR5) in this order.
The (CR1) command control unit 211 determines the copy start track number from the restore target track number (24-h) of the secondary volume (24-b) and the copy start track number (24-g) in the copy destination HDEV of the pair entry. Calculate the offset from (24-i). At this time, the difference between the restore target track number (24-h) and the copy start track number (24-g) is equal to the offset (24-i).
(CR2) The command control unit 211 calculates the restore destination track number (24-k) from the copy start track number (24-j) and the offset (24-i) in the copy source HDEV of the pair entry. At this time, the sum of the copy start track number (24-j) and the offset (24-i) is equal to the restore destination track number (24-k).
(CR3) The command control unit 211 acquires the difference state of the restore destination track number (24-k) from the offset (24-i) and the difference BM table 1911.
(CR4) If the difference state of the restore destination track number (24-k) is “unupdated”, the command control unit 211 does not perform the process of (CR5) and selects the restore target track (24-h). Increment (S2405) and perform processing from (CR1).
(CR5) If the difference state of the restore destination track number (24-k) is “updated”, the command control unit 211 stores the restore target track number (24-h) of the secondary volume (24-b). The track data is copied to the track of the restore destination track number (24-k) of the primary volume (24-a) (S2404).
Finally, a completion response to the differential restore command is sent to the MF host computer 102.

  Since the secondary volume (24-b) is a virtual volume configured from a virtual storage area, the operation of reading data from the copy source track is the same as the reading of the virtual volume described in the first embodiment. In the copy pair between the normal volume attribute HDEV 311 (primary volume) (24-a) and the virtual volume attribute HDEV 311 (secondary volume) (24-b), the secondary volume (24-b) to the primary volume is described above. This is the differential restore operation to (24-a).

  Further, as shown in FIG. 27, when receiving the reverse restore command from the management terminal 103, the command control unit 211 performs the differential restore from the secondary volume (24-b) to the primary volume (24-a), and then restores the primary volume and the secondary volume. The volume HDEV number may be exchanged to enable subsequent read or write access from the MF host 102. This operation is called reverse restore processing. Note that the differential restore operation step (S24 **) in FIG. 24 and the reverse restore operation step (S27 **) in FIG. 27 perform similar operations.

  Further, during the differential restore from the secondary volume (24-b) to the primary volume (24-a), the restore destination track number (24-k) or the restore target track number (24-h) When a read or write access is performed from the MF host 102, the command control unit 211 checks the difference state of the restore destination track number (24-k) described above. As a result, if the differential state of the restore destination track number (24-k) is “updated”, the restore destination track number (24-k), Alternatively, read or write access to the restore target track number (24-h) may be possible. This operation is called on-demand restore processing.

  As can be seen from the above description, even in the data copy processing from the normal volume to the virtual volume, it is possible to access the track mapping information in the hierarchical memory so that the owner right does not straddle.

  Next, the allocated track release processing in the HDEV 311 with the virtual volume attribute (when there is no record in the write target track when writing from the MF host computer 102 to the HDEV 311 with the virtual volume attribute) will be described with reference to FIG.

  When the command control unit 211 receives a write access command from the MF host computer 102 to the HDEV 311 with the virtual volume attribute, the operation of FIG. 25 is started (S2501). At this time, the command parameters are the HDEV number (25-a) of the primary volume, the track number (25-b) of the write target track, the write data, and the like. The command control unit 211 writes the data sent from the MF host computer 102 to the write target track according to the operation at the time of write access to the virtual volume described in the first embodiment (FIG. 17) (S2502).

If there is no record in the write target track due to write access from the MF host computer 102 (S2503: “Yes”), the command control unit 211 performs the following operations (RM1) to (RM5).
(RM1) The command control unit 211 acquires the track number (25-c) in the LDEV from the write target track number (25-b) and the TMT 1511.
(RM2) The command control unit 211 calculates the page number (25-d) in the LDEV and the offset (25-e) from the top track number in the page from the track number (25-c) and the page size.
(RM3) The command control unit 211 deletes (discards) the track entry (25-f) corresponding to the track number (25-b) of the write target track on the TMT 1511 (S2504).
(RM4) The command control unit 211 determines that the page number in the LDEV that can be calculated from the track number and page size in the LDEV out of all the track entries whose allocation determination information on the TMT 1511 is “allocated” is the page number It is checked whether one or more track entries (25-g) matching (25-d) exist other than the track entry (25-f).
In the processing of (RM5) (RM4), if there is no track entry (25-g) other than the track entry (25-f), the command control unit 211 corresponds to the page number (25-d) on the PMT 408. The page entry to be deleted is deleted (discarded) (S2505).

Next, a description will be given of garbage collection and allocated page release processing that accompany the allocated track release processing in the virtual volume. When one or more track entries (25-g) exist in addition to the track entry (25-f) in the above-described allocated track release processing (RM1) to (RM5), the following (GC1) and (GC2) ) Is executed.
(GC1) The track entry in which the track number in the LDEV is larger than the track number (25-d) and is the largest among the track entries whose assigned determination information on the TMT 1511 (FIG. 15) is “unassigned” (25-h) is acquired.
(GC2) The command control unit 211 registers the track number in the LDEV on the track entry (25-h) in the track entry (25-f), and sets the allocation determination information to “allocated”. In addition, the command control unit 211 deletes the track entry (25-h).
This rearrangement of track entries is called garbage collection.

  If the track entry (25-h) is deleted by the garbage collection process ((GC1) and (GC2)), the command control unit 211 releases the allocated track described above ((RM1) to (RM5)). The (RM4) and (RM5) processes in the above can be performed to release the allocated page. This process enables reclamation and garbage collection in units of tracks on the virtual volume, and enables efficient use of storage resources.

  Next, FIG. 26 shows the operation of ownership transfer processing of the virtual volume attribute HDEV 311 (when the command control unit 211 receives an ownership transfer instruction from the management terminal computer 103) according to the fourth embodiment of the present invention. explain. Command parameters include ownership transfer target HDEV number (26-a), ownership transfer source MPPK number (26-b), ownership transfer destination MPPK number (26-c), and the like.

  The command control unit 211 is registered in the TMT 1511 in the HDEV of the ownership transfer target HDEV number from the track entries cached in the LM 142 in the MPPK corresponding to the ownership transfer source MPPK number (26-b). All track entries are acquired and discarded from the LM 142.

  The command control unit 211 acquires the ownership entry (MPPK number) corresponding to the ownership transfer target HDEV number (26-a) on the management table 401 (FIG. 5) that manages the correspondence between the MPPK number and the HDEV number. . The MPPK number field 501 of the ownership entry is set to the ownership transfer destination MPPK number (26-c). The above is the ownership transfer processing of the virtual volume attribute HDEV 311. As a result, the owner right can be flexibly changed (moved), and the MP 141 to which the owner right has been delegated can access the track mapping information.

  Next, an access operation in units of files according to the fifth embodiment of the present invention will be described with reference to FIGS. FIG. 28 is a block diagram showing the overall configuration of the storage system, as in FIG. The difference from FIG. 1 is that the NFS / CIFS client 104 is connected to the network 111 via the communication line 117 and accesses the storage apparatus 101 in units of files.

  Next, FIG. 29 shows a file 290 exchanged between the NFS / CIFS client 104 and the storage apparatus 101. The file 290 includes a meta information part 291 and a data part 292. The meta information part 291 includes a file name 2911, a file type 2912, a file size 2913, an offset 2914 that is a start position of the meta information part 291 and the data part 292, and date information when the file is created, updated, and referenced (read). 2915. The meta information part 291 of the file 290 corresponds to the track control information in block unit access from the MF host computer 102.

  Next, the operation will be described by taking a write access request as an example. The NFS / CIFS client 104 requests the storage apparatus 101 to write the file 290 (issue a write command). The request is transmitted to the storage apparatus 101 via the network 111 and the communication line 115 connected to the communication line 117. In the storage apparatus 101, the NIC 125 receives the request, and the SWPK 126 transfers the request to the MPPK 123 in charge of processing.

  As described with reference to FIG. 17, the control information at the time of write access from the MF host computer 102 includes the HDEV number, the track number in the HDEV, and the track size. In the write access from the NFS / CIFS client 104, the information in the meta information unit 291 corresponds to the control information at the time of write access from the MF host computer 102. The unit of the data writing area is also adjusted to the page size (for example, 59,392 × 672 bytes).

  Through the above processing, not only host system access (block unit access) by the MF host computer 102 but also open system access (file unit access) by the NFS / CIFS client 104 is possible.

  As described in the first embodiment to the fifth embodiment, it is possible to access the track mapping information formed in the hierarchical memory of the distributed memory system so that the ownership does not straddle. Further, even in the data copy processing from the normal volume to the virtual volume, it is possible to access the track mapping information in the hierarchical memory so that the owner right does not straddle. Furthermore, reclamation and garbage collection can be performed in units of tracks on the virtual volume, and efficient storage resource operations can be performed. In addition, this invention is not limited to the said Example.

  The present invention can be applied to information processing apparatuses such as large computers, servers and personal computers, and information / video storage apparatuses such as storage systems and HDD recorders.

101 Multiple Storage Devices 102 MF (Main Frame) Host Computer 103 Management Terminal Computer 104 NFS / CIFS Client 111, 112 Network 113, 114, 115, 116 Communication Line 121 CHAPK (Channel Adapter PacKage)
122 DKAPK (DisK Adapter PacKage)
123 MPPK (Micro Processor PacKage)
124 CMPK (Cache Memory PacKage)
125 NIC (Network Interface Card)
126 SWPK (Switch PacKage)
131 Storage Controller 132, 314 Disk 141 MP (Micro Processor)
142 LM (Local Memory)
143 SM (Shared Memory)
144 CM (Cache Memory)
201 Microprogram 202 CM control information 203 Management table cache 211 Command control unit 212 RAID control unit 213 Ownership control unit 290 File 291 Meta information unit 292 Data unit 301, 3011, 3012 OS (Operating System)
311 Logical device for host (HDEV)
312 Logical device (LDEV)
313 Pool 401 MPPK number-HDEV number correspondence management table 402 HDEV number-LDEV number correspondence management table 403 HDEV number field-Volume size correspondence management table 404 HDEV number-volume attribute correspondence management table 405 HDEV number-disk number correspondence management table 406 LDEV Number-Pool Number-PMT (Page Mapping Table) Address Management Table 407 Pool Number-Disk Number Management Table 408 PMT (Page Mapping Table)
409 PMT management directory 410 TMT (Track Mapping Table) management directory 501 MPPK number field 502 HDEV number field 601 HDEV number field 602 LDEV number field 701 HDEV number field 702 Volume size field 801 HDEV number field 802 Volume attribute field 901 HDEV number field 902 Disk number field 1001 LDEV number field 1002 Pool number field 1003 PMT address field 1101 Pool number field 1102 Disk number field 1201 Page number field 1202 Page type field 1203 Page head address field in LDEV 1204 Disk number field 1205 Top of page address field 1206 page assigned judgment information field in the disk,
1301 LDEV number field 1302 PMT address field 1401 LDEV number field 1402 Initial data track number field in LDEV 1403 Next allocated track number field in LDEV 1511 TMT table 1501 Track number field in HDEV 1502 Track number field in LDEV 1503 Track allocated Judgment information field 1811 Copy BM table 1801 Track number field 1802 Copy status field 1911 Table for managing track number and update status (difference BM table)
1901 Track number field 1902 Update status field 2011 Pair management table 2001 Pair number field 2002 Copy source HDEV number field 2003 Copy source track range field 2004 Copy destination HDEV number field 2005 Copy destination track range field 2006 Copy BM table address field 2007 Difference BM table Address field 2911 File name 2912 File type 2913 File size 2914 Offset 2915 Date information

Claims (2)

In a storage system connected to a computer,
An interface unit connected to the computer;
A plurality of control units connected to the interface unit and configured by a plurality of microprocessors;
A shared memory shared and accessed by the plurality of control units;
A local memory shared and accessed by the microprocessor of the control unit;
A physical storage device for storing data used by the computer;
A first logical storage area composed of a set of tracks and assigned to the computer ;
A second logical storage area managed by a page composed of one or a plurality of tracks and assigned to the first logical storage area;
Providing a physical storage area composed of the physical storage device,
A first table for associating a track number of the first logical storage area with a track number of the second logical storage area and a track assignment state is arranged in the second logical storage area, and the first table is Assigning the first logical storage area to the second logical storage area,
A second table for associating a page of the second logical storage area with a page of the physical storage area and a page allocation state is arranged in the shared memory, and the second logical storage area is used by using the second table. Is assigned to the physical storage area,
When there is the access request from the computer to the first logical storage area, the access control only unit selected from the previous SL plurality of controller by referring to the first table and said second table Process the request,
When the access request is a data write request, the control unit performs data write to the write target track, and then determines whether there is a record storing the target data of the data write request in the write target track. When the record does not exist , the storage system is characterized in that the entry of the write target track is deleted from the first table and the allocation of the track is released. The storage system according to claim 1,
The control unit, after releasing the allocation to the write target track, refers to the first table , determines whether there is a track allocated other than the track in the page including the write target track, If there is no such track, the storage system is characterized in that the page entry is deleted from the second table and the allocation of the page is released.

Images